Predict after probit for different values

[Soeren Dallmeyer]

Hi,

I am running a simple probit model with Y as DV, X as a binary IV, and Z a vector for control variables.

probit Y X Z

After the estimation I want to look at the different predicted probabilities for both values of X. How can I do that?

Since I am interested in the marginal effects, I assume that the difference between those two predicted probabilities should give me the marginal effect, correct? I know that I could run the margins command. However this does work in my scenario for various reasons.

Thank you for your help.

[Carlo Lazzaro]

Soeren:
this recent Stata thread may be useful: https://www.statalist.org/forums/for...ch-observation

[Soeren Dallmeyer]

Hi Carlo,

Thank you very much. That was very helpful.